The impact of education and digitalization on female labour force participation in BRICS: an advanced panel data analysis

The study investigates how digitalization and education affect female labour force participation (FLFP) in BRICS countries (Brazil, Russia, India, China, South Africa), while also considering fertility and economic growth (GDP). FLFP is central to socioeconomic development, women’s empowerment, and household well-being, yet is shaped by mobility, segregation, and gender norms. Rapid digitalization—accelerated by COVID-19—has transformed work, enabling remote and platform-based jobs but also exposing digital divides. BRICS governments have promoted digitalization to spur productivity, competitiveness, and jobs. Given BRICS’ sizable population share (~42%), GDP share (~24%), and trade share (~16%) globally, and wide variation in FLFP and internet use, the bloc provides a salient context. The study poses three questions: (1) Does digitalization enhance FLFP in BRICS? (2) Does education increase FLFP? (3) Does higher GDP facilitate FLFP?

Prior studies generally find that ICT access and diffusion are associated with higher FLFP, though results can vary by context. Evidence from developing countries and regions (e.g., Africa, South Asia, Jordan, Latin America, India) suggests that internet and broadband penetration support women’s employment and participation, often via flexible work and improved job search; however, some cases (e.g., Saudi Arabia, Turkey) show null or adverse effects, potentially moderated by financial development or broader socioeconomic conditions. Education typically raises women’s labour force participation across diverse settings (Pakistan, Turkey, Botswana, Nigeria), though mismatches and constraints can limit employment gains despite rising education (e.g., Kerala, India). Fertility is commonly found to reduce FLFP in numerous cross-country and regional studies (G7, ASEAN, Middle East, Asia), with some evidence of time- or region-specific nuances. Regarding GDP and FLFP, many studies document a U-shaped relationship across development stages, while others find positive or insignificant links depending on context. The literature reveals limited, comprehensive analysis of the digitalization–FLFP nexus specifically for BRICS, motivating the present study.

Data and variables: Annual panel for BRICS (Brazil, Russia, India, China, South Africa), 1990–2020. Dependent variable: FLFP (female labour force participation, WDI). Explanatory variables: Digi (digitalization, measured by internet users as % of population), Edu (female secondary school enrolment, WDI), Fert (fertility rate), GDP (constant 2010 USD, WDI). All variables are log-transformed to reduce skewness. The expected signs: Digi (+), Edu (+), Fert (−), GDP (+). Econometric strategy: Given potential cross-sectional dependence (CSD), heterogeneity, and structural breaks, the study proceeds as follows: (1) Test for CSD using Pesaran (2015). (2) Assess stationarity using third-generation panel unit root tests accommodating CSD and breaks: Bai and Carrion-i-Silvestre (2009) and Pesaran (2007). (3) Examine cointegration with Westerlund and Edgerton (2008) and Banerjee and Carrion-i-Silvestre (2017), which handle heterogeneity, serial correlation, CSD, and structural breaks. (4) Estimate long-run and short-run relationships using the Cross-Sectionally Augmented ARDL (CS-ARDL) model, which augments regressions with cross-sectional averages of dependent and independent variables to address CSD and slope heterogeneity, and yields error-correction dynamics. (5) Conduct robustness checks using Augmented Mean Group (AMG; Eberhardt and Teal, 2010) and Common Correlated Effects Mean Group (CCEMG; Pesaran, 2006), both designed to accommodate unobserved common factors, CSD, heterogeneity, and structural breaks. Model: FLFP_it = β1i + β2i Digi_it + β3i Edu_it + β4i Fert_it + β5i GDP_it + φ_i + ε_it, with cross-sectional averages included per CS-ARDL specification. Long-run coefficients are derived from short-run parameters; the error-correction term (ECT) captures adjustment to equilibrium.

- Cross-sectional dependence: Significant CSD detected for all variables (e.g., FLFP: 21.020, p<0.001), motivating CSD-robust methods. - Stationarity: Bai and Carrion-i-Silvestre (2009) indicate non-stationarity at levels with stationarity at first differences; Pesaran (2007) suggests some stationarity at levels. Overall, variables treated as I(1). - Slope heterogeneity: Swamy-type tests (Pesaran and Yamagata, 2008) reject homogeneity, supporting heterogeneous panel methods. - Cointegration: Westerlund and Edgerton (2008) and Banerjee and Carrion-i-Silvestre (2017) confirm panel cointegration among FLFP, Digi, Edu, Fert, and GDP across the full sample and within countries. - Long-run CS-ARDL estimates (coefficients; all significant unless noted): • Digi: 0.261 (p<0.001) – higher internet penetration increases FLFP. • Edu: 0.361 (p<0.001) – higher female secondary enrolment raises FLFP. • Fert: −0.220 (p<0.05) – higher fertility reduces FLFP. • GDP: 0.199 (p<0.001) – economic growth increases FLFP. - Short-run CS-ARDL: Digitalization, education, and GDP positively associated with FLFP; fertility negatively associated; short-run magnitudes are smaller than long-run effects. The error-correction term ECT(−1) = −0.211 (p<0.001) indicates convergence to long-run equilibrium. - Robustness (AMG, CCEMG): Results corroborate that digitalization and education, as well as GDP, are positively related to FLFP, while fertility is negatively related, confirming CS-ARDL findings.

Findings directly address the research questions for BRICS: (1) Digitalization significantly increases FLFP, consistent with mechanisms such as improved job search, telework, and reduced transaction costs. (2) Female education has a strong positive effect on FLFP, reflecting higher employability and skills. (3) GDP growth is associated with higher FLFP, aligning with development-led expansion of employment opportunities. Fertility exerts a negative influence on women’s participation, consistent with caregiving constraints. The existence of long-run cointegration indicates stable relationships among these variables, and the smaller short-run effects suggest gradual adjustment dynamics typical for developing economies. Results align with much of the international literature while providing BRICS-specific evidence using advanced panel methods resilient to CSD, heterogeneity, and structural breaks.

This study offers the first comprehensive BRICS-focused empirical assessment of how digitalization and education affect female labour force participation, incorporating fertility and GDP and employing advanced panel econometrics. It documents robust, long-run positive effects of digitalization, education, and GDP on FLFP, and a negative effect of fertility. Short-run effects are directionally similar but smaller, with significant error-correction dynamics. Policy implications include: promoting digitalization through investment in ICT infrastructure and affordable internet access; expanding female education—especially ensuring completion of higher secondary education; and leveraging digital tools (e.g., remote work) to enhance women’s labour market participation. Future research could extend the model by incorporating additional macroeconomic factors (e.g., inflation, poverty, policy tools, technological innovation, remittances), apply the framework to other regional blocs, and compare across income groups to assess heterogeneity in effects.